The present invention is directed to the use of the maize Ac/Ds transposable elements in vertebrates, including fish, birds, and other animals including mammals and humans.
The publications and other materials used herein to illuminate the background of the invention, and in particular, cases to provide additional details respecting the practice, are incorporated by reference, and for convenience are referenced in the following text by author and date and are listed alphabetically by author in the appended bibliography.
Transgenic animals, including fish, provide an excellent vertebrate model for studying many facets of development, physiology and disease. A wide variety of fish may be utilized for this purpose. Exemplary fish include teleost fish, such as zebrafish (Danio rerio), medaka (Oryzias latipes), mummichog (Fundulus heteroclitus), killifish (Genus Fundulus), catfish (Genus Ictalurus), such as channel catfish; carp (Genus Cyprinus), such as common carp; and trout or salmon (e.g., Genus Salvelinus, Salmo, and Oncorhynchus). Zebrafish have become an established model for investigating many facets of development, physiology and disease.
Zebrafish are particularly useful for studying many facets of development, physiology and disease. They are small, develop ex utero, and have a short generation time. At 5 days of age each fish is a free swimming/feeding organism complete with most of the organ systems employed by mammals, such as heart, brain, blood, and pancreas. Within the last 10 years, mutant zebrafish lines isolated from large-scale mutagenesis screens have led to a greater understanding of vertebrate development (Driever et al., 1996; Haffter et al., 1996; Golling et al., 2002). Although these studies have shown that zebrafish mutants can serve as good models for human diseases, zebrafish have not been widely used in this capacity. So far the biggest limitation in zebrafish research has been determining the identity of causative genes disrupted from these mutagenesis screens, as the vast majority of mutants have been created using the chemical mutagen N-Ethyl-N-nitrosourea (ENU). Identification of ENU-derived point mutations requires laborious and slow positional cloning efforts. Insertional mutagenesis using retrovirus, on the other hand, is effective, and a gene mutated in this way can take as little as 2 weeks to identify with minimal resources (Golling et al., 2002). By creating a bank of retroviral insertions in zebrafish sperm, a library can be created where at least one insert into roughly every gene is housed in a small freezer space. Characterization and determination of the genomic location of all the inserts will make it possible to easily perform reverse genetics in the zebrafish by using the sperm sample with the mutation of interest to fertilize eggs in vitro. In this way, researchers could save a great deal of time and money by ordering mutations of interest instead of random screening. An even more powerful use of the library would lie in forward genetic screens.
To date, the number of cloned ENU mutants in zebrafish remains extraordinarily low considering the number of labs currently working on the hundreds of mutant lines. About 100 genes mutated with ENU have been published since the completion of two large-scale ENU mutagenesis screens in 1996. In fact, many of these genes were not identified using strictly positional cloning efforts. Rather, they were found by recognizing that certain mutant phenotypes were similar to known Drosophila or mouse mutants whose genes and pathways had already been decoded (Talbot and Hopkins, 2000). In these instances a “candidate gene” approach was taken whereby each gene in the presumed developmental pathway was examined individually for mutations to correctly isolate the disruption. While this method has proven extremely successful in cloning mutant genes, it has not led to the discovery of novel genes/pathways and has mainly recapitulated that which is known about development from other organisms.
A more successful approach to cloning mutant genes in zebrafish has been to use pseudo-typed mouse retroviruses as the mutagen. These retroviruses have the vesicular stomatitis virus G coat protein that allows infection of a broad range of host cells including zebrafish (Yee et al., 1994; Emi et al., 1991; Burns et al., 1993). The viral DNA inserts into the genome as a single-copy entity in a mostly random fashion without altering its junctional sequence, although retroviruses have been seen to prefer 5′ ends of genes as their insertion site (Vijaya et al., 1986; Rohdewohld et al., 1987; Mooslehner et al., 1990; Scherdin et al., 1990). Further, the retroviral insert serves as a molecular beacon, making it a rather simple process to link a mutant phenotype to a disrupted gene. One drawback is that the virus needs to be injected into zebrafish embryos at the 1000- to 2000-cell stage when the germ cells are still dividing (a necessary event for retrovirus integration). This requires more work up front than the traditional ENU mutagenesis method to create mutagenized founder fish. Since there are, on average, fewer retroviral insertions than ENU lesions per gamete, the mutagenic frequency of the retrovirus is less than that of chemical mutagens (1/1 ENU mutagenized F2 families produce a visible recessive mutation compared to 1/7 retroviral F2 families (Golling et al., 2002).
Recently, the retroviral method of mutagenesis has been used to conduct a large-scale developmental screen. That screen generated more than 500 mutants affecting zebrafish development; more than half of these disrupted genes were cloned (Golling et al., 2002; Amsternam et al., 1999). Results of this work have shown that all of the identified disrupted genes have homologues in human, but approximately 20% of these disrupted genes do not contain any obvious motifs or features that would allow one to classify the biochemical function of the resultant protein (Golling et al., 2002). In contrast, the genes so far cloned from ENU mutants show little degree of novelty.
One approach which has shown a great deal of promise for reverse genetics in zebrafish is the generation of gene knock-downs using morpholino based oligonucleotides (Heasman, 2002; Nasevicius and Ekker, 2000; Ekker, 2000). This technique relies on generating a short (24mer) morpholino oligonucleotide that is complimentary to the translation start site of the gene of interest (Summerton et al., 1997, Summerton and Weller, 1997). Injection of a morpholino oligonucleotide at the 1-2 cell stage inhibits translation of the endogenous target gene's mRNA. In this way many ENU derived mutant phenotypes have been phenocopied by injecting morpholino oligonucleotides specific to the mutated gene, establishing the proof in principle of the technique (Heasman, 2002; Nasevicius and Ekker, 2000; Ekker, 2000). These oligonucleotides have also been successfully used to examine unknown gene function of genes identified from in situ hybridization screens (Sakaguchi et al., 2001; Tsang et al., 2002).
While morpholino antisense technology is widely used in the zebrafish community it has some severe limitations. For instance, the window of opportunity for studying a knock-down of a favorite gene is only 2-3 days, thus limiting the technique primarily to the study of early development. Since the morpholino is not a stable, heritable element the amount of morpholino oligonucleotides in each cell is diminished by degradation and dilution with every round of division. Hence, if the desired gene is not expressed within this time period, which may presumably be the case with many disease genes, then this method will not work. The ability to perform suppressor or enhancer modifier screens would also be limited with morpholino oligonucleotides as each embryo in the screen would need to be injected with the oligonucleotide, a very time consuming effort.
Recent advances in producing high-titer retrovirus have greatly simplified its use in zebrafish. Some of the first retroviral constructs used for insertional mutagenesis in zebrafish generated injected (founder) fish that transmitted proviral integrations to only 5% of their F1 progeny (Lin et al., 1994). The number of inserts carried by these F1 fish was also low, usually one insert per gamete and fewer than 5 total insertions per germ-line (Lin et al., 1994). To do any meaningful mutagenesis screens with these constructs one would have to inject zebrafish embryos for a period of many years to generate 500,000 inserts. As seen in Chen et al. (2002), new retroviral vectors have significantly reduced this time frame. Now two people injecting retrovirus could make enough founder fish to harbor a half-million inserts in two months time (Chen et al., 2002). Furthermore, the efficiency of the retroviral system now allows more than 25 different insertions on average for each founder fish. These fish can easily be raised, tagged for individual identification system with a novel marking system, and stored in approximately 200 fish tanks, which is a small sized aquatic facility for a researcher.
Insertional mutagenesis is now the quickest method for cloning mutated genes in the zebrafish. Not only does the retroviral insert serve as a molecular tag for the disrupted gene, it also serves as a valuable marker to establish genetic linkage with the mutant phenotype. Chemical mutagenesis methods must rely on establishing tight linkage of the disrupted gene to a marker, thereby narrowing the genomic region of interest to a size small enough that it can be managed by sequencing of a BAC or PAC clone. The complication with chemical mutagenesis derives from the fact that there are hundreds of markers to test and they must be tested on a significant number (a few thousand) of recombinant fish to show that the linked marker does not segregate from the mutant locus. In screening for embryonic mutations, generating a few thousand mutant embryos is not usually a problem. However, if one were to look for adult phenotypes, such as those particular to certain diseases (diabetes, Parkinson's, obesity, etc.), it would take an inordinate amount of time, space, and resources to raise all of the required recombinants to find a linked locus. With insertional mutagenesis, linkage is established simply by running a Southern blot of the restriction enzyme digested DNA from the mutant fish and the pairs of adult fish that generated the mutants. Using a labeled portion of the retrovirus as a probe, one would expect all of the mutant fish and their parents to have one band migrating at the same molecular size, while pairs of fish that did not produce phenotypic clutches would not have the same insert (i.e., both fish would not have the linked band). In this way one would only have to look at a handful of affected fish to generate a probable lead on the disrupted gene.
The most time consuming aspect of cloning genes disrupted by insertional mutagenesis has been in cases where the genomic DNA flanking the retroviral insert does not contain exonic or known gene sequences. Usually when the candidate linked insertion is identified and its flanking sequence is cloned, it is no longer than a few kilobases in length due to the ineffectiveness of PCR to amplify larger fragments. In previous work, it was found that about one-third of the flanking DNA cloned in this way did not have useful information in that their sequences did not reveal any homology or identity in database searches (G. Golling, unpublished result). This will no longer be a problem as the first draft of the zebrafish genome will be completed soon and it should be possible to identify which gene is disrupted by sequencing less than 50 bases of DNA flanking the retroviral insert.
Reverse genetic approaches in mouse have provided many insights into human diseases. Researchers have been able to take a gene of interest and disrupt its expression by homologous recombination in ES cells, then re-introduce those cells back into an embryo to create targeted transgenic knock-outs. While this approach has not yet proven successful in zebrafish or other vertebrate model organisms other than the mouse and rat, another method, called target-selected mutagenesis, has been developed to create targeted disruptions in specific genes of interest. Target-selected mutagenesis is accomplished by first mutagenizing germ-line DNA of an organism and then using PCR to amplify the gene of interest. Sequencing the PCR product for comparison to the wild type gene then identifies samples containing mutations in the gene of interest. Wienholds et al. have recently used this technique in zebrafish to isolate mutations in the Rag1 gene. By sequencing two exons of Rag1 from a sperm library consisting of nearly 2700 randomly ENU mutagenized males, the researchers found 15 mutations, one of which was a premature stop codon (Wienholds et al., 2002). While the methodology does in fact work, it may be cumbersome for large-scale screening. To identify the single stop mutation approximately 12,500 sequencing reactions were carried out over a period of two months (Wienholds et al., 2002). Extrapolating this method to the entire zebrafish genome (40,000 genes) would take approximately 500 million sequencing reactions. A more efficient method would be to create a mutant sperm library from retroviral insertions. Since the retrovirus can serve as a tag for the disrupted locus many fewer sequencing reactions would be needed. The actual number of bases to be sequenced would also be smaller given the imminent completion of the zebrafish genome. Over a period of 4-5 months, three injectors could routinely generate the approximately 40,000 male founders necessary to harbor more than 1 million inserts. This would give an insert density on average of 1 every 1800 bp, essentially at least one insertion for every gene. Space requirements for housing the frozen library would be smaller than what is needed for most cDNA or genomic library arrays.
Forward genetics has been an invaluable approach in many model organisms; however, almost all of the forward and reverse genetics taking place in zebrafish are in the form of loss-of-function alleles. Forward genetic approaches based on the proposed retroviral zebrafish library would be enhanced with the addition of functional genetic elements. Retroviruses can provide additional functions aside from the obvious loss-of-function gene disruptions. For instance, in one large scale screen the predominant retroviral vector used had a gene-trapping cassette (Golling et al., 2002; Amsterdam et al., 1999). Of the founders injected with this construct, Chen et al. (2002) found there was at least one trapping event in the germ-line of each fish. While the trap vector itself did not prove to be more mutagenic than the previous non-trapping vectors (Chen et al., 2002), it did show the possibilities for other creative screens in the zebrafish that have previously been performed in Drosophila. Among these are enhancer trap constructs where fish could be screened for particular gene expression patterns, and promoter-containing vectors for mis-expression/over-expression screens. This latter approach has been used successfully in flies where Gal4 binding sites within a P-element drive expression of genes located downstream of the randomly located insert (Rorth et al., 1998; Hay et al., 1997). Transgenic fly lines expressing Gal4 protein in a controlled manner, for example under the control of an eye specific enhancer or tetracycline operator, would cause the gene downstream of the P-element to be mis-expressed according to the researcher's preference. A function such as this would provide a valuable resource for studying the function of genes that do not display obvious loss-of-function phenotype. It is estimated that such genes constitute more than two-thirds of genes in flies, worm, and yeast (Sulston et al., 1992; Dujon et al., 1994; Miklos and Rubin, 1996). It is likely that an even higher percentage of vertebrate genes have no obvious loss of function phenotype. These genes are often biologically important. For example, although loss of either NPY and/or AGRP function in mice display no detectable abnormalities (Qian et al., 2002; Erickson et al., 1996), NPY and AGRP have been found to play a key role in regulation of food-intake by gain-of-function studies (Levine and Morley, 1984); Clark et al., 1984; Graham et al., 1997; Ollmann et al., 1997). Furthermore, their pathways have been drug discovery targets for obesity and diabetes in several pharmaceutical and biotech companies (Halford, 2001).
It is estimated that half of the genes identified in the first draft of the human genome have no function ascribed to them (Lander et al., 2001; Venter et al, 2001). The ability to rapidly examine the biological roles of these unknown genes is the goal of many research institutions and pharmaceutical companies. Currently, the best vertebrate model organism for conducting these genetic studies is the mouse. Homologous recombination and random retroviral mutagenesis have made the mouse a viable resource for functional genomics research. However, several drawbacks inherent in the mouse's biology have stymied the development of quick, large-scale approaches towards gene and gene function identification. These include significantly large space requirements, small litter sizes, development in utero, high cost of breeding and maintenance, and vast regulatory/animal handling requirements. In zebrafish nearly the opposite is the case concerning these issues. Zebrafish have the added advantage of being transparent through most of their development, allowing easy visualization of the morphology and function of internal organs by light microscopy using a variety of techniques involving fluorescent, luminescent or colorimetric labeling. Thus, developing zebrafish as a forward genetic model would significantly enhance the understanding of vertebrate gene and protein function.
Recently an effort has been made by at least two companies to generate a zebrafish sperm library and, in fact, one loss-of-function gene has been published from this library (Nasevicius and Ekker, 2000). Since the library was made from fish mutagenized with ENU, several thousand PCR and sequencing reactions were required to isolate the lone mutant. Using retroviral insertional mutagens would vastly improve the speed with which a zebrafish library could be generated and then used by researchers. Stored as sperm samples and/or as approximately 20,000 or fewer fish, possibly as few as 2,000 fish or fewer, this insertional library occupies little space. The sperm samples can easily be reconstituted into living fish for studies of gene and/or protein function. The fish can easily be reproduced for studies of gene and/or protein function. A large repository of cloned retroviral mutations in the zebrafish would be a valuable resource for the study of the function of a specific gene of interest, for screening the library for phenotypes relevant to disease so as to identify putative drug targets, for screening the library for fish that do not respond to significant drugs, toxins or other chemicals so as to identify the gene and protein that are the site(s) of action of such compounds, and for screening for compounds that might alter the expression or activity of known disease genes, or other genes/proteins of interest.
DNA transposons are mobile elements that can move from one position in a genome to another. Naturally, transposons play roles in evolution as a result of their movements within and between genomes. Geneticists have used transposons as tools for both gene delivery and insertional mutagenesis or gene tagging in lower animals (Shapiro, 1992) but not, until recently, in vertebrates. Transposons are relatively simple genetic systems, consisting of some genetic sequence bounded by inverted terminal repeats and a transposase enzyme that acts to cut the transposon out of one source of DNA and paste it into another DNA sequence (Plasterk, 1993). Autonomous transposons carry the transposase gene inside the transposon whereas non-autonomous transposons require another source of transposase for their mobilization.
One well known transposable element is the maize Ac/Ds element (Shure et al., 1983; Fedoroff et al., 1983; Pohlman et al., 1984). Maize Ac/Ds elements can transpose in a wide variety of plant species (Osborne and Baker, 1995). Moreover, successful Ds transposition catalyzed by modified transposase was demonstrated in Saccharomyces cerevisiae (Weil and Kunze, 2000), implying that plant-specific proteins were not essential for transposition. Transposition in heterologous animal hosts has been reported for a number of other transposons e.g. mariner element from Drosophila mauritiana in zebrafish (Fadool et al., 1998), nematode Caenorhabditis elegans Tc3 element (Raz et al., 1998) in zebrafish, synthetic transposon Sleeping Beauty in mammalians and zebrafish (Horie et al, 2001; Davidson et al, 2003; Balciunas et al., 2004), Tol2 from Oryzias latipes in zebrafish, Xenopus and mouse (Kawakami et al., 2000; Kawakami et al., 2004; Kawakami and Noda 2004) etc. However, none of the known transposons has yet been demonstrated to transpose in both plants and animals.
The use of heterologous transposons has been a powerful tool for genetic research in a number of model species (Parinov et al., 1999; Spradling et al., 1995). The advantage of using heterologous elements is that there is no transposase in a new host once inserted a genomic copy of non-autonomous element is immobile but can be mobilized if transposase is delivered into the cells.
Methods for introducing DNA into a cell are known. These include, but are not limited to, DNA condensing reagents such as calcium phosphate, polyethylene glycol, and the like), lipid-containing reagents, such as liposomes, multi-lamellar vesicles, and the like, and virus-mediated strategies. These methods all have their limitations. For example, there are size constraints associated with DNA condensing reagents and virus-mediated strategies. Further, the amount of nucleic acid that can be introduced into a cell is limited in virus strategies. Not all methods facilitate integration of the delivered nucleic acid into cellular nucleic acid and while DNA condensing methods and lipid-containing reagents are relatively easy to prepare, the incorporation of nucleic acid into viral vectors can be labor intensive. Moreover, virus-mediated strategies can be cell-type or tissue-type specific and the use of virus-mediated strategies can create immunologic problems when used in vivo.
There remains a need for new methods for introducing DNA into a cell, particularly methods that promote the efficient integration of nucleic acid fragments of varying sizes into the nucleic acid of a cell, particularly the integration of DNA into the genome of a cell. There also remains a need to develop a vertebrate (e.g., a zebrafish, mouse, etc.) insertional mutation library that could be used to screen for genetic defects, to study genes of interest, to screen for drugs useful for treating or preventing a disease condition associated with a gene of interest.